Response of soil extracellular enzyme activity and stoichiometry to short-term warming and phosphorus addition in desert steppe

Study site

The experiment was conducted on a desert steppe in Eastern Yanchi County (37°04′–38°10′N, 106°03′–107°04′E), Ningxia Hui Autonomous Region, Northwest China. It is located on the southwest edge of the Mu Us Desert, which has a semiarid continental monsoon climate. According to meteorological data from the Yanchi Meteorological Station, the average annual temperature is 8.8 °C, and the average annual precipitation is 298.15 mm, with the majority of the precipitation falling between July and September (1980–2021). The soil type is classified as Arenosol (IUSS, 2015), with 4.12 g kg⁻¹ of soil organic matter, 0.40 g kg⁻¹ of total nitrogen, 0.32 g kg⁻¹ of total P, 2.30 mg kg⁻¹ of available P, and pH of 8.57. The dominant species in the region are Agropyron mogolicum, Lespedeza potaninii, Caragana korshinskii, Stipa bungeana, and Polygala tenuifolia Willd.

Experimental design

The experimental warming device was modified based on the meteorological data from the study site from 1980 to 2021 (Fig. S2) as well as previous research (Ma et al., 2019; Ma et al., 2021). Open top chambers (OTCs) were used to passively increase the temperature in this investigation. Stainless steel and high-transmittance glass material (5 mm thick) were used to construct a regular octagonal prism structure, with a substrate area of 5.6 m² and a vertical height of 1.8 m. These were permanently installed in the sample plot to avoid disruption. The air and soil temperatures at 15 cm above and below ground were automatically recorded every half-hour using HOBO MX2302A and HOBO MX2201 (Onset Computer Corporation, Bourne, MA, USA) data loggers, respectively.

Soil sampling and measurements

We sampled soils on August 25, 2022. Three soil columns (5 cm in diameter) were collected randomly from the surface soil (0–15 cm in depth) in each subplot. The soil samples were mixed after removing gravel and litter, and then homogenized and sieved through a 2-mm mesh sieve. The composite soil samples were then divided into two parts, the first subsample was air-dried to measure soil organic carbon (SOC), total nitrogen (TN), total phosphorus (TP), and available phosphorus (AP). The second subsample was stored at 4 °C to measure dissolved organic carbon (DOC), ammonium nitrogen (NH₄⁺–N), nitrate nitrogen (NO₃⁻–N), microbial biomass (MBC, MBN and MBP), and soil extracellular enzyme activities.

Soil moisture content (SMC) was determined by oven-drying the soil at 105 °C for 48 h. A PHS-3E glass pH electrode (Leici, Shanghai, China) was used to measure the pH of the soil in a suspension of air-dried soil and distilled water (1:5, w/v). Soil organic carbon (SOC) and total nitrogen (TN) were determined using the potassium dichromate external heating method and H₂SO₄-H₂O₂ digestion with the Kjeldahl method. Total phosphorus (TP) was determined by HClO₄-H₂SO₄ digestion with the molybdenum antimony anti-colorimetric method (Nelson & Sommers, 1996). A TOC analyzer was used to measure dissolved organic carbon (Vario TOC, Elementar, Hanau, Germany). AN (NH₄⁺–N and NO₃⁻–N) was determined using the KCl extraction method with a continuous flow analyzer (AutoAnalyzer3, Bran Luebbe, Germany). Available phosphorus (AP) was determined by NaHCO₃ extraction using the molybdenum antimony anti-colorimetric method; and microbial biomass (MBC, MBN, MBP) was extracted and analyzed by chloroform fumigation (Vance, Brookes & Jenkinson, 1987).

Soil extracellular enzyme extraction and vector analysis

Soil C- (β-1,4-glucosidase, BG), N-(leucine aminopeptidase, LAP, and β-1,4-N-acetylglucosa minidase, NAG), and P- acquiring enzyme (alkaline phosphatase, ALP) activities were determined using a modified standard fluorescence technique (Saiya-Cork, Sinsabaugh & Zak, 2002; German et al., 2011). Soil extracellular enzyme activity (EEA) was expressed in units of nmol g⁻¹ h⁻¹ (Sinsabaugh et al., 2008; Sinsabaugh, Hill & Shah, 2009; Waring, Weintraub & Sinsabaugh, 2014). The ratios of C, N, and P acquisition enzymes were calculated using the following formulae to determine soil extracellular enzyme stoichiometry (EES), respectively:

Soil e C: N = ln (BG) /ln (LAP + NAG)

Soil e C: P = ln (BG) /ln (ALP)

Soil e N: P = ln (LAP + NAG) /ln (ALP)

Vector analysis of soil extracellular enzyme activity was used to identify microbial nutrient limitations (Hill et al., 2012; Jiang, Zhu & Zeng, 2022). Vector length and vector angle were calculated by the following formulae, respectively (Moorhead et al., 2013; Moorhead et al., 2016), with limitations on N and P shown by vector angles of less than 45° and greater than 45°, respectively, and greater C limitation is shown by a relatively longer vector length:

Relative C limitation = Vector Length = $\sqrt{{\left[ln\left(\mathrm{BG}\right)/ln\left(\mathrm{LAP}+\mathrm{NAG}\right)\right]}^2+{\left[ln\left(\mathrm{BG}\right)/ln\left(\mathrm{ALP}\right)\right]}^2}$

N/P limitation =Vector angle(°) = Degrees (ATAN2((ln(BG)/ln(ALP), (ln(BG)/ln(LAP+NAG)))

Statistical analyses

A two-way analysis of variance using a split-plot design was carried out to investigate the effects of warming, P addition, and their interaction on soil extracellular enzyme activities and stoichiometries in desert steppe. One-way ANOVA was used to compare the differences in soil properties, soil extracellular enzyme activities, and stoichiometries of the different treatments. Significant differences were tested using Duncan’s multiple comparison tests (P < 0.05), and the data of soil enzyme activity was transformed using a natural logarithm before statistical analysis. Before performing an ANOVA, all data were tested for normality of the residuals and homogeneity of variances in SPSS 25.0 for Windows (SPSS Inc., Chicago, Illinois, USA). Pearson correlation analysis was used to determine the association between soil properties, microbial biomass, and stoichiometries. Mantel test was used to explore the correlation between C, N, and P acquisition enzyme activities and soil environmental factors. Redundancy analysis (RDA) was used to explore the corresponding relationships of soil properties, microbial biomass, and extracellular enzyme activity in Canoco 5.0. For the graphics, R 4.2.3 and Origin 2022 (Origin Laboratory Corporation, Northampton, MA, USA) were applied.

Soil temperature and soil moisture content

OTCs had the anticipated warming effect in all temperature-increased treatment plots (Fig. 1), with an average increase of 1.15 °C in the air temperature 15 cm aboveground and an average increase of 0.65 °C in the soil temperature 15 cm belowground. Warming significantly decreased soil moisture content (SMC) by 13.67%, while P addition significantly increased SMC by 11.20% and 19.52% at P₅ and P₁₀, respectively. Soil moisture content was not significantly affected by warming × P addition (Table 1).

Soil properties

Warming had no significant effects on SOC, TN, and TP (Table 1). P addition did not affect SOC, but significantly increased TP and decreased TN. Warming × P addition increased SOC and TP. WP₁₀ decreased TN by 4.48%, while WP₅ increased it by 13.43%. Warming and P addition did not significantly affect SOC: TN, but decreased SOC: TP. Warming × P addition increased SOC: TP by 1.43% at WP₅ and decreased by 22.59% at WP₁₀. TN: TP decreased by 2.58% and 29.38% at WP₅ and WP₁₀, respectively.

Soil pH was not significantly affected by warming, P addition, and their interaction (Table 1). Warming and P addition both individually decreased DOC, but their interaction significantly increased DOC by 12.22% and 15.06% at WP₅ and WP₁₀, respectively. Warming and warming × P addition significantly increased AN. P addition decreased AN by 6.67% at P₅ and did not change at P₁₀. P addition had a considerable impact on available P, and the higher the rate of P addition, the greater the available P content. However, warming had no discernible influence on available P. Warming and P addition decreased DOC: AN. Warming × P addition decreased DOC: AN by 33.05% at WP₅ and no significant effect at WP₁₀. P addition and warming × P addition had a significant negative impact on DOC: AP and AN: AP.

Soil microbial biomass and their stoichiometry

Warming and P addition did not significantly affect MBC and MBN, but their interaction significantly decreased both MBC and MBN (Figs. 2A and 2B). Warming reduced MBP by 9.01%. P addition and warming × P addition considerably increased MBP (Fig. 2C). Warming and P addition, individually, significantly decreased MBC: MBN, MBC: MBP, and MBN: MBP. Warming × P addition increased MBC: MBN (Fig. 2D), and substantially decreased MBC: MBP (Fig. 2E) and MBN: MBP (Fig. 2F).

Soil enzyme activity and their stoichiometry

Soil BG (C-acquiring enzyme) was significantly decreased by warming, P addition, and their interaction (Fig. 3A). As for warming × P addition, BG at WP₅ was higher than at WP₁₀. LAP+NAG (N-acquiring enzyme) was significantly decreased by P addition (Fig. 3B), but not significantly affected by warming or warming × P addition. ALP (P-acquiring enzyme) was significantly reduced by warming, P addition, and warming × P addition (Fig. 3C). Warming and P addition reduced ln (BG): ln (LAP+NAG) (Fig. 3D), while warming × P addition increased ln (BG): ln (LAP+NAG) by 4.10% at WP₅, and decreased it by 13.93% at WP₁₀. Warming × P addition increased ln (BG): ln (ALP) at WP₅ (Fig. 3E), but there were no significant changes observed on ln (BG): ln (ALP) from other treatments. Warming, P addition, and their interaction raised ln (LAP+NAG): ln (ALP) (Fig. 3F). A positive correlation between soil C-, N-, and P-acquiring activity was also discovered (Fig.  4).

Vector analysis

All the treatments were over the 1:1 line (Fig. 5A), which showed that microbial nutrients were severely phosphorus limited. Vector lengths of P₀ and WP₅ were relatively greater than other treatments. Warming and P addition decreased vector lengths (Figs. 5B and 5C). Warming × P addition increased vector length by 11.64% at WP₅ and decreased vector length by 7.53% at WP₁₀. All the vector angles were greater than 45° , and vector angles were reduced by warming, P addition, and their interaction (Fig. 5D).

Relationships between soil properties and soil extracellular enzyme activities

Pearson correlation analysis showed a significant relationship between AP and DOC: AP, AN: AP, MBN: MBP, and MBC: MBP. Soil moisture content was significantly related to DOC: AP and AN: AP. The Mantel correlation analysis showed that soil total nutrients and stoichiometries had no significant effects on soil extracellular enzyme activities and stoichiometries. However, a positive link between soil C and P acquisition enzyme activities and soil available nutrients, microbial biomass, and their stoichiometry was observed. Soil N- acquiring enzyme activity was closely related to DOC: AN, DOC: AP, and MBP. (Fig. 6). The connection between soil available nutrients, soil extracellular enzyme activities and stoichiometries, and vectors were then examined (Fig. 7). The results revealed that the important factors affecting soil C-, N-, and P-acquiring activities and stoichiometries were AN, MBC: MBP, and MBN: MBP. There was a significant positive correlation between AN and soil e C: P; Soil e N: P was positively correlated with AP; MBC was positively correlated with soil e C: N; Vector length was positively correlated with AN, while vector angle was significantly negatively correlated with AP (Fig. 7).

1

Short-term warming

Supporting our hypotheses, soil C-, N-, and P-acquiring enzymes were decreased by warming (Fig. 3). Warming increased soil available nitrogen contents (ammonium nitrogen and nitrate nitrogen, Table 1) and decreased microbial biomass C and N (Fig. 2), but there were no significant effects on the soil total nutrients and stoichiometries (Table 1). There was a significant positive correlation between soil extracellular enzyme activities (EEAs) and soil available nutrients, microbial biomass, and their stoichiometry (Figs. 6 and 7). These results indicated that increased temperature affected the secretion of soil enzymes by altering the available nutrients. This aligns with the results of previous studies (Cui et al., 2021; Li et al., 2022b). Further research on the soil microbial communities of desert steppes is required to better understand the exact microbiological mechanisms by which microbial biomass control soil extracellular enzyme activity.

Soil moisture content (SMC) may also affect soil extracellular enzyme activity (Allison & Treseder, 2008; Li et al., 2022a). Warming considerably decreased SMC in our study (Table 1), but had no significant impact on soil C-, N-, and P-acquiring activities (Figs. 6 and 7). It is suggested that soil microorganisms may adapt to dry, infertile soil conditions by adjusting soil extracellular enzyme activity in response to their demands and the surrounding environment (Wang et al., 2017). In our study, warming raised ln (BG): ln (ALP) and ln (LAP+NAG): ln (ALP), and decreased ln (BG): ln (LAP+NAG) (Fig. 3), vector length, and vector angle (Fig. 5). This is likely because warming reduced the soil microbial biomass C and N, and consequently influenced the secretion of C- and N-acquiring enzymes to adapt to environmental changes (Turner & Wright, 2014).

Seasons also have a significant effect on soil extracellular enzyme activity (Machmuller et al., 2016), which may indicate variations in substrate accessibility and environmental circumstances across different seasons. A long-term warming experiment is needed to observe the larger impact of seasonality on soil extracellular enzyme activity. In this study, short-term warming changed soil microbial activities and stoichiometries in desert steppe.

Phosphorus addition

Soil phosphatases typically showed higher activity in P-limited ecosystems, and soil C- and N-acquiring enzyme activities reflect microbial demand for energy and nutrients (Turner & Wright, 2014). In our study, P addition decreased soil extracellular enzyme activities (Fig. 3), which had a great impact on P acquisition enzyme activity (Marklein & Houlton, 2012; Shi et al., 2021). P addition significantly promoted soil total P and available P contents while concurrently reducing DOC and AN (Table 1). P addition significantly increased MBP and decreased MBC and MBN (Fig. 2), suggesting that soil microorganisms may preferentially fix P because low inorganic P availability induces rapid microbial P uptake (Bünemann et al., 2012). Our correlation analysis demonstrated a negative correlation between soil C-, P- acquiring enzyme activity and available P, MBP (Fig. 6). This relationship is most likely explained by P addition increasing plant growth and nutrient uptake while reducing the number of nutrients available for the synthesis of soil enzymes (Wang, Wang & Liu, 2008). These results are contrary to the results of some previously published research, which can likely be explained by differing factors such as the rate of P addition, type of ecosystem, and study site (Colvan, Syers & O’Donnell, 2001; Tian et al., 2016; Wang et al., 2020b).

In terms of soil extracellular enzyme stoichiometry, P addition decreased ln (BG): ln (LAP+NAG), and increased ln (BG): ln (ALP) and ln (LAP+NAG): ln (ALP) (Fig. 3). This finding illustrates that P addition decreased microbial demand for carbon and soil microbial demand for alkaline phosphatase production costs in phosphorus-limited soils. This may also impact nutrient cycling, plant productivity, and microbial community composition (Peñuelas et al., 2013; Luo et al., 2019). P addition decreased vector length and vector angle (Fig. 5), demonstrating that P addition can alleviate both the soil microbial carbon and phosphorus limitation (Wang et al., 2022). This result confirmed our hypothesis that soil microbial nutrient limitation of the desert steppe is closely related to soil available P.

The interaction effects of short-term warming and P addition

Warming × P addition significantly reduced soil C-, N- and P-acquiring enzyme activities (Fig. 3). As for C-acquiring enzyme activity, WP₅ was higher than WP₁₀, which may be related to the significant increase in soil DOC and AN. This, in turn, may have caused the low level of P₅ addition rate to secrete more C and N acquisition enzymes at higher temperatures (Fig. 2). At WP₁₀, there was less soil available C and N, and fewer C- acquiring enzymes were secreted by the microorganisms. WP₁₀ was consistently higher than WP₅ throughout the changes in N-, and P-acquiring enzyme activities, suggesting that the amount of available phosphorus may be a vital factor (Table 1). P-acquiring enzymes were most significantly impacted by warming × P addition (Fig. 3), indicating a negative feedback relationship between high available phosphorus content and soil extracellular enzyme activities (Allison et al., 2007). The positive correlation between BG, LAP+NAG, and ALP indicates that the microbial acquisition to C, N, and P were altered due to the changes in soil substrate availability (Sinsabaugh, Hill & Shah, 2009). Warming × P addition resulted in a reduction of ln (BG): ln (ALP) and ln (LAP+NAG): ln (ALP), an increase in ln (BG): ln (LAP+NAG) at WP₅, and a decline in ln (BG): ln (LAP+NAG) at WP₁₀ (Fig. 3). The primary factors that influenced soil extracellular enzyme activities and stoichiometries were AN, MBC: MBP, and MBN: MBP. These results show that the interaction of warming and P addition had an impact on soil physicochemical properties, which altered the stoichiometry of soil extracellular enzymes (Zheng et al., 2015; Zhang et al., 2018).

Additionally, the natural log ratio of soil C-, N-, P- acquiring enzymes in this study was 1.2:1:1.3 (Table S1), which differs from the global mean scale of 1:1:1 (Sinsabaugh, Hill & Shah, 2009), demonstrating that the desert steppe in the study region was largely C and P limited. This is partially consistent with the findings of a previous study demonstrating that a desert steppe was severely P-limited in the temperate grassland of northern China (Peng & Wang, 2016). The ratio of soil C: N: P acquisition enzymes adjusted to 1.3:1:1.2 at WP₅. A low phosphorus addition rate caused both soil microbial carbon and phosphorus limitation in the region, while the addition of P₁₀ alleviated it (Table S1). Previous research revealed that adding P may boost soil microbial nitrogen demand, and lead to an increase in N-acquiring enzyme activity (Wang et al., 2020b). However, this did not occur in our study, and this finding may need to be confirmed by further research. Our vector analysis revealed that relative carbon limitation increased at WP₅ and decreased at WP₁₀. A higher level of P addition rate (P₁₀) was beneficial to alleviating P limitation in the region. Warming × P addition changed the soil available nutrients and their stoichiometry, while soil nutrient stoichiometry influenced microbial C, N, and P metabolism by regulating soil elemental balance (Cui et al., 2019b). In this process, the secretion of soil microbial enzymes in P-deficient areas of the desert steppe was stimulated to adapt to nutrient limitation (Schimel, Balser & Wallenstein, 2007). According to our findings, the key influencing factors of microbial nutrient restriction in desert steppe are soil available nutrients, microbial biomass, and stoichiometries.